How Research Ethics Committees Can Adapt to the Rise of Generative AI: Strengthening Governance, Oversight, and Ethical Review

Generative artificial intelligence (AI) has rapidly transformed the research landscape across universities, research institutions, healthcare organizations, and industry laboratories. From generating text and images to assisting with coding, data analysis, and literature reviews, tools such as large language models (LLMs) are increasingly integrated into research workflows. While these technologies offer unprecedented opportunities for innovation, they also introduce complex ethical challenges that traditional research governance structures were not designed to address.

Research Ethics Committees (RECs), also known as Institutional Review Boards (IRBs) in some jurisdictions, play a critical role in ensuring that research involving human participants adheres to ethical principles such as respect for persons, beneficence, justice, and transparency. However, the emergence of generative AI has expanded the scope of ethical considerations beyond conventional concerns. Researchers can now create synthetic datasets, generate participant-facing materials, automate decision-making processes, and deploy AI-powered interventions that may affect participants in unforeseen ways.

As generative AI becomes embedded in academic research, ethics committees must evolve their review processes, expertise, and governance frameworks to address emerging risks. Effective adaptation requires understanding how AI changes research methodologies, identifying new ethical challenges, updating review protocols, and building institutional capacity for responsible AI oversight.

The Impact of Generative AI on Research Ethics

Generative AI is reshaping the way research is conducted across disciplines. Researchers increasingly use AI systems to draft research proposals, summarize literature, analyze qualitative data, create educational materials, and support decision-making processes. While these capabilities can improve efficiency and productivity, they also raise questions about accountability, transparency, data integrity, and participant protection.

One of the most significant impacts of generative AI is the changing nature of authorship and intellectual contribution. Researchers may rely heavily on AI-generated content while still claiming sole authorship of publications. Ethics committees must therefore consider whether AI-assisted research processes are adequately disclosed and whether researchers understand the limitations of AI-generated outputs.

Generative AI also introduces challenges related to reproducibility. Traditional scientific methods rely on transparent and repeatable procedures. However, AI-generated outputs may vary depending on prompts, model updates, and system configurations. This variability can complicate efforts to verify research findings and assess methodological rigor.

Another important concern is participant interaction. AI systems are increasingly used in surveys, interviews, educational interventions, and healthcare applications. Participants may unknowingly interact with AI-generated content, creating ethical concerns regarding informed consent, transparency, and autonomy.

Furthermore, generative AI systems are often trained on vast datasets that may contain copyrighted material, personal information, or biased content. Researchers using these systems may inadvertently introduce ethical issues into their studies if the underlying AI models produce discriminatory, inaccurate, or misleading outputs.

As a result, research ethics committees must broaden their evaluation criteria  to include AI-specific risks while maintaining established ethical principles.

New Ethical Risks Introduced by AI Tools

Generative AI technologies present a range of novel ethical risks that require careful scrutiny during ethics review AI studies.

Bias and Discrimination

AI models learn from historical data that may contain societal biases. Consequently, generative AI systems can perpetuate or amplify discrimination based on race, gender, ethnicity, socioeconomic status, disability, or other protected characteristics.

Research projects using AI-generated outputs may unknowingly expose participants to biased information or unfair treatment. Ethics committees should require researchers to assess potential bias risks and implement mitigation strategies before study approval.

Privacy and Confidentiality

Researchers frequently use generative AI tools to process sensitive information. However, entering participant data into external AI platforms may create privacy risks if information is stored, reused, or accessed by third parties.

Ethics committees must evaluate whether researchers have established appropriate safeguards for data handling, anonymization, and confidentiality. Special attention should be given to studies involving health data, educational records, or other sensitive information.

Misinformation and Hallucinations

Generative AI systems sometimes produce inaccurate or fabricated information, commonly referred to as hallucinations. If researchers rely on AI-generated outputs without verification, misinformation may influence study findings or participant experiences.

For example, AI-generated educational materials or health information could inadvertently provide incorrect guidance. Ethics committees should require validation procedures to ensure the accuracy of AI-generated content used in research.

Lack of Transparency

Many advanced AI systems operate as black boxes, making it difficult to understand how outputs are generated. This lack of transparency can hinder accountability and make ethical evaluation more challenging.

Researchers should be encouraged to document AI tools used, explain their role in the research process, and disclose known limitations. Ethics committees may require additional documentation to assess whether AI use aligns with ethical standards.

Synthetic Data Risks

Generative AI enables the creation of synthetic datasets that mimic real-world data. While synthetic data can support privacy-preserving research, it may still contain identifiable patterns or reinforce existing biases.

Ethics committees should evaluate whether synthetic data adequately protects privacy and whether researchers have assessed potential risks associated with its use.

Manipulation and Deception

AI-generated content can be highly persuasive and personalized. In research settings, this capability raises concerns about participant manipulation, undue influence, and deception.

Studies involving AI-generated communications, virtual assistants, or conversational agents should be carefully reviewed to ensure participants are treated fairly and informed appropriately about AI involvement.

Updating Ethics Review Processes

To address emerging risks, ethics committees must modernize their review procedures and integrate AI-specific considerations into existing frameworks.

Expanding Ethics Application Requirements

Research proposals involving generative AI should include detailed information regarding:

• AI tools and platforms being used
• Intended purpose of AI within the study
• Data sources and training considerations
• Privacy and security safeguards
• Bias assessment procedures
• Human oversight mechanisms
• Plans for monitoring AI performance

Including these elements enables ethics committees to conduct more comprehensive evaluations.

AI Risk Assessment Frameworks

Ethics committees can adopt structured risk assessment models that categorize AI applications according to potential impact.

Low-risk applications might include administrative support or literature summarization, while high-risk applications may involve healthcare decisions, participant profiling, or automated interventions.

Risk-based review processes allow committees to allocate resources efficiently while ensuring greater scrutiny for higher-risk studies.

Enhanced Informed Consent Requirements

Traditional consent forms may not adequately address AI-related concerns. Participants should understand:

• Whether AI systems are involved in the study
• How AI-generated outputs will be used
• Potential limitations of AI technologies
• Privacy implications associated with AI processing
• Measures taken to protect participant rights

Clear disclosure promotes transparency and supports informed decision-making.

Ongoing Monitoring and Post-Approval Oversight

Unlike conventional research tools, AI systems can evolve over time through updates and retraining. Ethics review should therefore extend beyond initial approval.

Committees may require periodic reporting on AI performance, emerging risks, and mitigation efforts throughout the study lifecycle.

Continuous oversight enables timely identification and management of unforeseen ethical concerns.

Multidisciplinary Review Approaches

AI-related research often involves technical, legal, social, and ethical dimensions. Ethics committees should consider involving experts from multiple disciplines when reviewing complex AI studies.

This collaborative approach improves the quality of ethical assessments and strengthens responsible AI oversight.

Capacity Building for Ethics Committees

Successfully governing generative AI requires significant investment in knowledge development and institutional capacity.

Enhancing AI Literacy

Many ethics committee members were trained before the emergence of advanced generative AI  technologies. Consequently, they may lack the technical knowledge necessary to evaluate AI-related risks effectively.

Institutions should provide ongoing education covering:

• Machine learning fundamentals
• Generative AI capabilities and limitations
• Algorithmic bias
• AI governance frameworks
• Data protection requirements
• Emerging regulatory developments

Improved AI literacy enhances the committee’s ability to make informed decisions.

Recruiting AI Expertise

Ethics committees may benefit from including members with expertise in:

• Artificial intelligence
• Data science
• Cybersecurity
• Digital ethics
• Privacy law
• Human-computer interaction

Specialized expertise can help committees navigate complex technical issues and strengthen governance processes.

Developing AI Ethics Guidelines

Institutions should create clear policies outlining expectations for researchers using AI technologies.

Guidelines may address:

• Transparency requirements
• Data management practices
• Bias mitigation procedures
• Human oversight obligations
• Reporting responsibilities
• Accountability mechanisms

Institution-specific guidance promotes consistency and supports compliance.

Knowledge Sharing Networks

Research institutions can establish collaborative networks that facilitate the exchange of best practices, case studies, and lessons learned regarding generative AI governance.

Such networks enable ethics committees to remain informed about emerging challenges and evolving standards.

Scenario-Based Training

Practical exercises involving realistic AI research scenarios can help committee members develop confidence in evaluating AI-related proposals.

Scenario-based learning encourages critical thinking and improves decision-making skills in complex ethical situations.

Recommendations for Effective Oversight

As generative AI continues to evolve, research ethics committees must adopt proactive strategies to ensure effective governance.

Implement Risk-Based Governance Models

Not all AI applications present the same level of ethical concern. Committees should prioritize oversight based on the potential impact of AI systems on participants, society, and research integrity.

Risk-based governance enables efficient resource allocation while maintaining robust protections.

Require Transparency and Documentation

Researchers should provide comprehensive documentation regarding AI systems used in their studies, including limitations, assumptions, and intended functions.

Transparency supports accountability and facilitates ethical review.

Promote Human Oversight

Generative AI  should complement rather than replace human judgment in research activities.

Committees should encourage researchers to maintain meaningful human involvement in decision-making processes, particularly in high-risk applications.

Strengthen Data Governance

Effective oversight requires strong data governance practices, including:

• Data minimization
• Secure storage
• Access controls
• Anonymization procedures
• Compliance with privacy regulations

Robust governance protects participant rights and promotes trust.

Encourage Ethical Impact Assessments

Researchers should evaluate potential social, ethical, and legal consequences before deploying AI systems.

Ethical impact assessments help identify risks early and support responsible innovation.

Foster a Culture of Responsible Innovation

Ethics committees should move beyond compliance-focused approaches and encourage researchers to view ethics as an integral component of scientific excellence.

A culture of responsibility promotes innovation that aligns with societal values and public interests.

Monitor Regulatory Developments

The regulatory environment surrounding AI is evolving rapidly. Ethics committees should remain informed about international standards, government policies, and emerging best practices to ensure governance frameworks remain relevant.

Conclusion

The rise of generative AI presents both extraordinary opportunities and significant ethical challenges for the research community. Traditional ethics review frameworks remain valuable, but they must evolve to address the unique risks associated with AI-generated content, automated decision-making, synthetic data, algorithmic bias, privacy concerns, and transparency limitations.

Research ethics committees occupy a central position in ensuring that generative AI research ethics principles is upheld across academic and scientific institutions. By updating ethics review processes, enhancing AI literacy, recruiting multidisciplinary expertise, strengthening governance mechanisms, and implementing continuous oversight, committees can effectively manage emerging risks while supporting responsible innovation.

As AI technologies become increasingly integrated into research activities, proactive adaptation will be essential. Institutions that invest in modernized ethics governance today will be better positioned to protect participants, maintain public trust, and ensure that generative AI advances contribute positively to society.